U.S.-China AI crisis management talks focus on balancing competition for supremacy and safety

[Washington, Beijing Jiji] How can we balance the increasingly intense competition in artificial intelligence (AI) development with safety? Ahead of the U.S.-China summit meeting expected on the 24th, controversy is spreading around the world over ``AI crisis management'' to prevent AI from going out of control or being misused. This was triggered by a threat theory advocated by the head of a leading American AI company. According to British media, the two countries are expected to hold ministerial-level talks on AI on the 20th, and the focus will be on whether the two superpowers competing for high-tech supremacy can come to terms.

U.S. NVIDIA CEO says new regulations are "not needed" over talk of slowing down AI development

The person who started it was Dario Amodei, CEO of the American AI startup Anthropic. He warned that AI could take over the entire internet and advocated slowing down the pace of development and testing the safety of cutting-edge models. OpenAI CEO Sam Altman and American businessman Elon Musk also agreed.

The background to this is wariness of China, which is hotly chasing the US in AI. Mr. Amodei emphasized safety and called for restrictions on exports of semiconductors to China and crackdowns on the unauthorized use of AI models. In response, China objected, saying that the move was "inciting threats, conflict, and vicious competition" (Ministry of Foreign Affairs).

Europe aims to become a "third pole" that sets itself apart from the United States and China. Germany points out that the participation of the United States and China in AI crisis management is essential, and Spain calls for an "international agreement." The idea is to use common rules to minimize runaway and misuse while continuing development, likening it to international treaties that allow the peaceful uses of nuclear energy while preventing the proliferation of nuclear weapons.

However, the US and China are determined to accelerate AI development. U.S. President Donald Trump asserted, ``It's a hoax that AI will take over the world and destroy humanity.'' Chinese President Xi Jinping also positions it as a ``strategic tool to seize the lead in the global science and technology competition.'' Each side is concerned that slowing its own development will benefit the other.

There is a huge gap between the United States and China when it comes to creating rules to deal with the dangers of AI. "While slowing development could buy time for regulatory discussions, it is highly unlikely that China will join a global agreement," said David Sachs, a special adviser on AI in the Trump administration. China views the U.S.-led rules as "containment" for its own country.

The AI ​​debate has shifted from corporate-level security measures to national crisis management. If the leaders of the United States and China agree on mutual reporting of serious accidents and common safety inspections, it will serve as a "guardrail" to prevent the development competition from overheating. On the other hand, if crisis management is used as a tool to protect one's own country's advantage, the creation of common rules will be delayed. Can AI safety be separated from the struggle for supremacy? Dialogue between the two countries in the run-up to the summit will be a test of this.